Pixinsight

[johnrt]

Quite right Olly, there are wonderful images produced in PI without the need for any other applications, and I have no quarrel at all with that, or the fantastic imagers like Harry Page & Barry Wilson that make them, I just find some of the statements from the team behind PI arrogant and just plain wrong.

"Suum cuique!!"

I think the LHE tool actually has a slider in it called "Amount" that lets you choose the amount of the original and modified image to blend together in the resultant image, what is this if not a PS layer and opacity option from the stone age?

[swag72]

34 minutes ago, ollypenrice said:

Painting in Photoshop, Harry? Can you give me an example? I've heard this said before but I don't think I do it and I haven't seen anyone do it here, unless I'm misunderstanding the term......

I would say that my MCM experiments where I created a full colour version of an image from JUST Ha data would be considered as painting for sure ...... But I was very clear in the fact that I had done that and that it was nothing more than an artistic representation. I never tried to pass it off as anything other than art. 

[Gina]

I look upon astro image processing as a combination of science and art.  I was educated in science and had a career in it but I'm also an artist as a hobby/pastime - I enjoy both

[ollypenrice]

4 hours ago, johnrt said:

Quite right Olly, there are wonderful images produced in PI without the need for any other applications, and I have no quarrel at all with that, or the fantastic imagers like Harry Page & Barry Wilson that make them, I just find some of the statements from the team behind PI arrogant and just plain wrong.

"Suum cuique!!"

I think the LHE tool actually has a slider in it called "Amount" that lets you choose the amount of the original and modified image to blend together in the resultant image, what is this if not a PS layer and opacity option from the stone age?

Yes. The problem with the slider is that it remains global and the joy of Layers is that you can be selective by using the eraser. I've just opined that Gnomus' excellent Pacman would benifit from LHE used as a Ps Layer. I only ever use it this way because I never want it applied universally. I guess the PI gurus feel that if you don't do something to the whole of an image you are cheating. But you can do it to part of an image provided you use masks. This sounds like equivocation to me! (Like the monks who were not allowed to handle money so they wore gloves.)

Olly

[ollypenrice]

1 hour ago, Gina said:

I look upon astro image processing as a combination of science and art.  I was educated in science and had a career in it but I'm also an artist as a hobby/pastime - I enjoy both

I'm currently getting into trouble with Steve Swamp Thing for regarding it as primarily a craft!

Olly

[PatrickGilliland]

7 hours ago, Ikonnikov said:

I would broadly agree with your sentiments Paddy with a couple of caveats maybe.

Deconvolution is supposed to recover details lost to blurring and therefore generate an image closer to 'reality' and as such is considered a restorative process at least when the point spread function is empirically determined for the system (which is convenient in astronomy since we have plenty of point spread functions (stars) in our images). Not to say it doesn't have its issues and subjectivity; the software has to fit the stars properly to its model(s) and also being an iterative process there is the question of how many iterations is enough? Also useful restoration only occurs in image areas with sufficiently high snr so some method must be used to designate and protect low signal areas to prevent generation of artefacts. Needless to say the process is used professionally in astronomy and microscopy to improve image resolution.

Sharpening imho is a greyer area; contrast-limited adaptive histogram equalization (the method behind LHE) is also used professionally in image processing to enhance the definition of structure edges which are (in theory at least) only informed by the data in the original image and the contrast limitation helps to prevent over sharpening of lower signal areas. But again there is still subjectivity as to how much to limit the contrast and indeed in PI I think a luminance mask of some kind is essential to keep the noise down when using LHE.

So I would say the results of decon and LHE if used carefully can be largely reflective of the original data. But I would say since we have to adapt an optically imperfect linear image with features that vary massively in their brightness into one in which can be viewed on a monitor and appreciated by the human eye there can be countless numbers of different final images produced from the same original data which are all in their own way reflective of the original data. But beyond this I fully agree that star masks, selective sharpening of interesting features etc (independent of their signal intensity) which generate  images which are clearly no longer 'true' to the original data have a place in making the aesthetics of a great astronomical image.

Just my thoughts...

Paul

 

The bottom line (which seems to have been missed in later comments) is that if any one application was the only way forward and worked perfectly we would all be using that application but all producing the same exquisite image.  

On decon will concede that one - still subjective based on your inputs and interpretation but no image is ever in perfect focus, how close you bring it or exceed that point is still user skill.  On LHE  (shock horror) sometimes i might run two iterations and we all have our preferences on the settings.  Likewise if i run decon, HDMRT then LHE that can be very different to decon, LHE then HDRMT etc All of these factors make the final result artificial, but no less a good representation of the target.

Detracting though from the missed point - this being; there is no perfect application so whichever application is used the skill is that of the user to create a version that is pleasant.  We all start with different sets of imperfect data and manipulate that.  This is the skill.  Some programs may have tools and settings that are 'true to the data' but even this is the application designers interpretation of the 'rules' and if all our data is different every time it is run, even on the same target it will generate slightly different results.  It is the human intervention that inspects all of this and keeps the output of the applications on the straight and narrow.  Without the tools which are plentiful and varied we could not do but we chose which and when and this is, for me the skill.  This skill is that of the individual.

The flip side all gets a bit heavy, if we claim there is science involved you may need to read "Astrophysics of Gaseous Nebulae and Active Galactic Nuclei" before you next look at M42  

The tools are great - the users are the defining element in the production of excellent images.

Paddy 

[Barry-Wilson]

2 hours ago, PatrickGilliland said:

Detracting though from the missed point - this being; there is no perfect application so whichever application is used the skill is that of the user to create a version that is pleasant.  We all start with different sets of imperfect data and manipulate that.  This is the skill.  Some programs may have tools and settings that are 'true to the data' but even this is the application designers interpretation of the 'rules' and if all our data is different every time it is run, even on the same target it will generate slightly different results.  It is the human intervention that inspects all of this and keeps the output of the applications on the straight and narrow.  Without the tools which are plentiful and varied we could not do but we chose which and when and this is, for me at least anyway.

The flip side all gets a bit heavy, if we claim there is science involved you may need to read "Astrophysics of Gaseous Nebulae and Active Galactic Nuclei" before you next look M42  

The tools are great - the users are the defining element in the production of excellent images.

Paddy 

Here! Here!

This is an interesting thread and I have tried (in vain) to read only!  I have stated before that I do no believe in a PI vs PS dichotomy.  I don't view either as superior and certainly don't subscribe to the PI-is-truth doctrine and wholeheartedly agree that its terminology can be off putting. We have all seen both masterful and woeful images from PI, PS, Startools, GIMP, Fitswork etc etc - the software is a tool only and they broadly aim to achieve the same thing.  You simply use what you see fit to process your image how you envision it and strive to master the techniques over a lifetime of practise!  Good luck to us all whatever our software choices.

I have recently downloaded an interesting new entrant to the graphics software market, developed over the last couple of years or so I believe - correct me if I'm wrong - "Affinity".  It appears very PS-like to my untrained eye.  I think it is for both mac and PC, but am not sure (I use a mac), https://affinity.serif.com/en-gb/ .  Anybody used this in anger?

[ChrisLX200]

Axe or saw - both cut wood, different techniques but same end result - but only for a skilled user   

My problem with PI is that it's much harder to use well  than PS, and many users rely on tutorial walk-throughs which they apply to their own image data - often inappropriately. Way too many PI-processed images simply shriek "PI" ! at first glance which is not a fault of the program per se but simply reflects the requirement for a high degree of familiarity with how the proggram works in order to get the best out of it. It's staggering the number of galaxy images I see which appear to have been 'drawn around' to mask out where discrete differential development takes place only to leave an obvious false edge with bright galaxy and black background! Of course this is no fault of the program, it doesn't have to look like this, but the sheer difficulty of understanding the program and what all those parameters really mean (and the useful ranges of numbers that should/could be used with them) IS the main reason why it's difficult to use.

In all cases step one is to understand your own data and what it contains, step two is understanding all the tools PI offers and how they modulate the data and to develop a strategy for tackling it, step three is knowing [predicting/calculating] the parameter values to use with each tool, and step four is using those tools in the necessary sequence (with the appropriate parameters set) to perform the operation(s) to give the desired result. If what comes out the far end is a total surprise then you definitely have a problem with PI! With PS you at least get the satisfaction of instant and global visual feedback on how the parameters you're using are affecting the image and in that respect at least is easier to understand - you can develop a 'feel' for the data more easily I think and how far you can push it.

ChrisH

[Barry-Wilson]

Chris - I agree.  It is only with practise that recently I feel my images are escaping the 'PI' look.  PI certainly takes some deciphering and its seeming complexity is a barrier to its use.

The only reason why I haven't branched out to use PS too is the dread of facing another learning curve.  I have downloaded a trial of PS in the past and having been steeped in PI for years now PS did seem surprisingly alien and I genuinely had no clue where to start.  Adobe's rental scheme puts me off too.  Hence an interest in Affinity and I would be interested to hear if anyone has given this software a whirl.

[ChrisLX200]

@Barry - I've ordered that recent book by Warren Keller 'Inside Pixinsight' which I'm hoping will help my understanding of how it works, at the moment I can only use a few of the tools with confidence so I'm not really getting my money's worth out of it!  Pixinsight's capabilities are not in doubt - just my ability to use it which is pathetic

ChrisH

[Barry-Wilson]

As a point of interest, does anyone know what software professional astronomers use to process their data for 'art' purposes, eg HST images?

[harry page]

3 minutes ago, Barry-Wilson said:

As a point of interest, does anyone know what software professional astronomers use to process their data for 'art' purposes, eg HST images?

Hi

I do know that some of the hubble images are done with photoshop , but some of the chandra x ray observatory images are done with pixinsight

I did not process any of the Chandra stuff ( not allowed ) but was lucky enough to give some advice

Harry

[ollypenrice]

Perfect sanity emerges, yet again, on SGL. I'll endorse the observations of Paddy, Barry, Chris and Harry very happily.

Olly

[harry page]

Its a good Job I am sane here as my wife thinks I am Mad

 

Harry

[ollypenrice]

9 minutes ago, harry page said:

Its a good Job I am sane here as my wife thinks I am Mad

 

Harry

That really is interesting because, despite my preference for Photoshop, my wife also thinks I'm mad. No matter how many times I point out to her that I'm guiding at less than 0.5 arcsecs, and maintaining the steadiest course through life that anyone could wish for, she persists in this slanderous delusion. Honestly, it's no wonder I drink blood in the moonlight.

Olly